Adaptive programs enhance medical image classification with explainable evidence
Evidence Field: Adaptive Covering Programs for Medical Imaging
Artificial IntelligenceComputer Vision and Pattern RecognitionMachine Learning
Summary
Medical image classifiers usually tell us what they see but not how they made their decisions. The authors present Evidence Field, a method that not only classifies images but also gives a detailed, step-by-step explanation of how it arrived at its conclusion by breaking down the image into parts and combining evidence. This approach lets others independently check and reconstruct the decision process. They tested it on real medical images from different hospitals, showing it helps understand predictions better.
What this means in practice
- •For medical imaging engineers: Enable interpretable classification by producing reconstructable spatial evidence in medical image diagnosis workflows.
- •For histopathology informatics teams: Support multi-center evaluation of image analysis tools with patient-separated training and transparent, reproducible decision models.
Authors
Isarar Siddique, Raiyan Siddique, Arshi Niyaz, Jamyang Ugyen Tshomo, Atish Raza
Abstract
Medical image classifiers typically return predictions without a native contract for independently reconstructing their spatial decision computation. Evidence Field is a trainable spatial-program architecture that jointly produces class scores, selected covering programs and signed evidence. Its encoder combines absolute and image-relative response fields with associative spatial moments and learned characteristic sums. A typed classifier scores individual regions and parent-conditioned multiplicative groups; a covering recursion selects scale, partition and composition. We establish exact one-fold coverage, optimality within the registered grammar, a bound relating smooth training to hard inference, signed-score conservation and class-centred offset invariance. Direct contractions and their adjoint specify arithmetic and storage scaling on two- and three-dimensional grids. Native output records bind inputs, parameters, supports, types and contextual dependencies to individual and ensemble predictions. In a separately reported histopathology application, patient-separated training and evaluation across held-out centres provide empirical evidence for the architecture. Trained mechanism and corrected-loss comparisons distinguish predictive effects from implementation correctness; independent readers reconstruct descriptors, routes, scores and probability mixtures. The principal contribution is a coupled representation, decision and evidence architecture whose selected computation can be reproduced independently.